Research On The Application Of Multi-agent Technology To Supply Chain Management

Supply chain management (SCM), a new management system, has been a hottopic since the recent years. For there is no systematic theory frames for SCM atpresent, the discoveries of methodologies and practical measures are still in great needof significant efforts. With the development of information technology and World WideWeb, a number of new methods are coined, with agent being featured with theindependence, adaptation and communication capabilities. In the view of systemdynamics, this dissertation aims at the research on the application of agent technologyin artificial intelligence to the foundation, communication, coordination and simulationof supply chain system in order to provide a practical way for SCM. Significant resultsand useful conclusions have been achieved in the following aspects.Firstly, the disadvantages of the traditional models on SCM are discussed, andmulti-agent system (MAS) theory in distributed artificial intelligence is proposed. Amulti-agent model is introduced for SCM system, and the interdependent relationshipof all agent components and their functions are described. The MAS is characterizedwith distribution, open, adaptability and robust, so it can serve as a useful tool for thedescription of SCM. The MAS can be used to effectively model, optimize, implementand control the work of a supply chain.Secondly, the communication mechanisms among agents in MAS are investigated,and agent communication language (ACL) is studied. By making comparison ofKnowledge Query and Manipulation Language and FIPA ACL, the tendency for ACLin future is explored. Communication is a required way to acquire information forcoordination and negotiation. ACL has been a cornerstone for the development ofsystems of communicating agents, and simultaneously they have been the subject ofintensive standardization efforts.Thirdly, the coordination for MAS and the negotiation model are studied focusingon three coordination methods and control-and data-driving models. Negotiation isinstrumental for the intercommunication among the agents in MAS. Differentnegotiation methods are required for different purposes, so how to select a suitablenegotiation model is essential for the design of MAS.Finally, the software reuse technology and SCM are integrated, andsoftware-developing technology is applied to the model design and simulation framefor a multi-agent supply chain. Simulation is an effective way for analyzing andassessing a supply chain development and management in detail during the course ofits reconstruction. Under this frame, it is convenient and timesaving for a supply chainmodel development based on a software library.